Future communication infrastructure and technical challenges in the medical field Luis Loyola CTO Allm Inc.

Communication Platforms in Healthcare 2 Share patient information instantly and easily. ● High communication security ● High standards of patient data protection ● Following clinical protocols and workflows ● Integrating with Electronic Health Record (EHR) systems

Join: Agile communication among clinical professionals ● A communication app for the telemedicine among medical professionals ● The first app registered as a medical device and listed in public health system in Japan 3

Join chat and REST system architecture 4

Join DICOM integration 5

The Growing Need for Communication Platforms in HC ● Increase in patient volume: The population is growing and aging, and with it the demand for healthcare services is also increasing. This is putting pressure on healthcare systems to find more efficient ways to communicate with patients. ● Advancements in technology: The rise of telemedicine, electronic health records, and mobile health apps has increased the need for effective communication platforms to connect patients, healthcare providers, and other stakeholders. ● Cost savings: Implementing communication platforms can help healthcare systems save money by reducing the need for face-to-face visits, and by streamlining administrative tasks. ● Quality of care: Effective communication can improve the quality of care by allowing healthcare providers to collaborate more effectively and by providing patients with timely and accurate information. ● Patient engagement: By providing patients with easy access to information and resources, communication platforms can help to engage patients in their own care and improve their overall health outcomes. ● Remote care: The need for remote care has been increased during the pandemic, communication platforms can play a big role in providing care to patients who can not travel or have mobility issues. Overall, communication platforms have the potential to improve the efficiency, effectiveness, and accessibility of healthcare, which is why their usage is increasing. 6

The Growing Need for Communication Platforms in HC ● Increase in patient volume (aging population) ● Advancements in technology (telemedicine, EHR and mobile health apps) ● Cost savings (less face-to-face visits, and streamline of administrative tasks) ● Quality of care (allowing healthcare providers to collaborate more effectively and by providing patients with timely and accurate information) ● Patient engagement (easy access to information and resources engage patients in their own care and improve their overall health outcomes) ● Remote care (big role during the pandemic) ● AI Chatbots like ChatGPT Overall, communication platforms have the potential to improve the efficiency, effectiveness, and accessibility of healthcare, which is why their usage is increasing. 7

Improving patient outcomes with Time Stamps Stroke case AMI case Diabetic Foot case ER case 8

Improving patient outcomes with Time Stamps 9 Brain Stroke Heart Attack Emergency Network Diabetes

Integration with EHR, PHR and IoT sensors 10 Doctor to Doctor Communication apps (eg Join) Personal Health Record apps (eg MySOS, Apple Health) Electronic Health Record Systems (eg Human Bridge, TrackCare) FHIR-compatible Clinical Data Storage and Sharing (Patient Folder) IoT Sensors (Blood Pressure monitor, EKG, Eye Fundus camera, EEG, Glycemic Measurement, etc)

What is FHIR? ● Fast Healthcare Interoperability Resources (FHIR) RESTful API for health data ● HL7 standard with long history – HL7v2: internal hospital data (released in 1989) – HL7v3: Health Information Exchange (HIE) (released in 2003) – CDA/CCD: Clinical Document Architecture / Continuity of Care Document (released in 2005) – FHIR: Simplified REST implementation of CDA (leverages RESTful web services and open web technologies such as XML, JSON, and RDF) (released in 2014) 11

Current R&D projects ● Automated Clinical Trial Matching using AI ● Alert notifications based on Patient Health Data for PHR applications like MySOS ● AI Chatbots ● Automated text tagging using NLP 12

Current R&D projects ● Automated Clinical Trial Matching using AI ● Alert notifications based on Patient Health Data for PHR applications like MySOS ● Chatbots ● Automated text tagging using NLP 13

Automated Clinical Matching using ML 14 Patient Elegibility Criteria Deep Learning Model Model Extract Features from Patient Data Patient Database Patient Data in Multiple Formats Patient Matching Score ● Saves time to recruit patients ● Mitigate adverse events ● Highly reliable matching with accuracy ● Improved Risk Identification

Alert notifications based on PHR Data 15 Patient Health Data Deep Learning Model Model Extract Features from Patient Data Patient Database Patient Data in Multiple Formats Health Score Personalized alerts, warnings, recommendations ● Predictive Analytics: likelihood of developing a particular disease or the risk of a health complication ● Doctor/specialist visit and exam recommendations ● Lifestyle recommendations

CHATBOTS 16

Chatbots ● Triaging patients: Chatbots can be used to triage patients by asking them initial questions about their symptoms and medical history, and then directing them to the appropriate healthcare provider or service. ● Providing information: Chatbots can be programmed to provide patients with information about their condition, treatment options, and recovery. ● Scheduling appointments: Chatbots can be used to schedule appointments with healthcare providers, and to remind patients about upcoming appointments. ● Managing medication: Chatbots can be used to remind patients to take their medication, and to provide information about dosage and side effects. ● Follow-up care: Chatbots can be used to provide follow-up care to patients, such as checking in on their recovery and providing additional information or resources. ● Telemedicine: Chatbots can be used to provide remote consultations with healthcare providers, which can be particularly useful in rural or underserved areas. ● Research: Chatbots can be used to aid in clinical research by gathering data from patients and providing personalized information based on that data. Overall, chatbots can be a valuable tool in clinical communication systems, as they can improve efficiency, reduce costs, and enhance the patient experience. 17

Chatbot sample from Allm EMEA GmbH, “Jacob” 18 Status Quo ● Every hospital has different workflows ● -Every change in the application applies to all users / hospitals ● -But not every hospital may want a specific feature ● -New features in Join need to be implemented in 3 platforms plus server side + plus design etc.

Jacob’s benefits 19 ● Only one implementation (tiny docker container) necessary with a few lines of code is necessary for a chatbot ● Chatbot can be adopted to every customer needs ● Chatbot can do everything a "normal" user could do but may be too time consuming to do (especially during emergency) ● -"Jacob" appears as a friendly helper in Join

Jacob Technically 20 ● Small server (docker container) uses the existing Join API to access information shared with a Join user ● Bot can listen to new messages, dicoms etc. and act, i.e. send message, create case, send timestamps, transfer information to other platforms (e.g., HIS) ● This even works with end-to-end encryption Join Rest API Join Rest API

Jacob’s Example 1: Create Patient Cases 21 ● -JoinTriage message does not create a new patient case in Join ● -Users do not want to enter information twice ● -Chatbot listens for JoinTriage messages and creates a new case with all information provided by the message ●

Jacob’s Example 2: Real-Time Stats of Intensive Care Beds 22 ● -Chatbot extracts information from DIVI (German intensive bed registry) webpage and posts it every day into a chatgroup ● -Physicians do not need to actively look up info ● ●

Jacob’s Example 3: HIS integration 23 ● Emergency registers patient in IVENA Health (web-based system to send emergency services in Germany and Austria) ● Bot polls new patients every n seconds ● If it is a stroke patient, send it to the chat group in Join ● Create a Join case for the patient ● Track timestamps (additional timestamps via JoinForms) ● Send CSV file to MHH HIS with timestamps ● Doctors use CSV file (or PDF) for review / research ● ● ●

Jacob’s Example 3: HIS integration 24 Em ergency Case Registered Poll Relevant Cases Send Info Create Group Register Time Stamps Send M H H Patient ID

Jacob’s Example 4: Patient recruiting for Clinical Trials 25 ● Apps for Clinical Trials like Join Trials in Europe and Enroll in Japan ● Upon the fulfillment of a basic matching criteria, the chatbot asks the clinicians further questions in order to figure out whether that specific patient is a good candidate for Clinical Trial or not ● Frequently the recruiting of patients are randomized and distributed into two groups: the ones who receive the active treatment and the others who receive the placebo. ● The digitilization of this process helps specialists save their valuable time and make the enrollment process smoother. ● After the patient is selected, Jacob continue asking further questions during the active treatment in order to provide relevant follow-up information to the doctors and specialists. ● ● ●

Jacob’s Example 5: Customer Support and Sales Boosting 26 ● ChatGPT, BARD, etc. will become a commodity in a few years time. ● Product and commercial oriented bots that can clearly answer the inquiries, respond to complaints by customers, or even introduce new products or services to the current customers can become a very powerful tool for commercial teams. ● ●

Automated Text Tagging using NLP 27

Automated Text Tagging using Natural Language Processing JOIN REST API TEXT ANALYTICS JOIN REST API JOIN REST SERVER

Automated Text Tagging using NLP X Case Timeline Clinical Indication FHIR FHIR

Automated Text Tagging using NLP X Case Timeline Clinical Case Presentation FHIR FHIR

Automated Text Tagging using NLP X Case Timeline Medical Image Interpretation FHIR FHIR

Automated Text Tagging using NLP X Case Timeline Clinical Procedure FHIR FHIR

Automated Text Tagging using NLP X Case Timeline Clinical Indication FHIR FHIR

Conclusions ● Communication platforms for clinical use are becoming increasingly important as healthcare organizations move towards more digitized and connected systems. ● Functionalities and tools based on Natural Language Processing and Artificial Intelligence will play a significant role in their user adoption and interoperability during the next coming years. ● Connection with IoT devices, AI-powered Image Analytics, and EHR systems will also be highly relevant to create powerful tech ecosystems in healthcare. ● Other big challenges not covered in this presentation are Data Security and Privacy as well as Compliance with local and global regulations. Overall, clinical communication platforms have the potential to significantly improve patient care and streamline healthcare operations. 34

Luis Loyola loyola@allm.inc ありがとうございます、GRACIAS! 35